Preamp difference : if it's not the frequency, not the slew rate, and not the harmonics, what is it ?

[abbey road d enfer]

I just wonder which record / reply medium is going to cover 20Hz-80kHz range (at a price point which is viable).

DVD is capable of 24bit at 192kHz SR in stereo.
How many people are interested?

 

[RoadrunnerOZ]

Cubase has its own separate Control Room setup but Tools doesn’t. This seems a simple way of doing it:
https://www.production-expert.com/production-expert-1/how-to-create-a-listen-bus-in-pro-tools

 

[abbey road d enfer]

All tracks sent to the master as well as to a buss, pre-fader

Pre-fader to mix also?

 

[Mihobro]

Just my two cents,
On panning, hard left or hard right are the only positions of zero phase , Center in stereo is kinda there but still subject to anomolies of stereo speakers room acoustics etc. That is why theatre sound developed the center speaker. to increase the clarity of dialog. So there are few real choices for pure punchy reproduction until you get to the bottom frequencies where the two speakers function together as a single array. (That will depend on the spacing of the speaker pair.) So to pan outside of the hard left and right creates "fuzz" in the mid and high frequencies, and the real question is how much fuzz do you want. In a mix you can use that to contrast with the stuff that is in your face clear. Futher, adding of time delays choruses, Haas effect etc. can make mid and hi stuff clearer, but it is at expense of low and mid low frequencies which will add comb filtering that ruin any phase linearity. To add to the complication, people hear things differently. Some are able to isolate sounds, hearing around all the fuzz. To others it is just one big din. Others have one ear that hears things different from the other so they will favor one side over the other. So trying to come to a consensus is sometimes impossible. So unless there are individual speakers in all over the place - super Atmos surround? (Has anyone heard playback in George Massenbergs's Blackbird room) We make the best decision in the given situation with the given listeners.

 

[JohnRoberts]

Just my two cents,
On panning, hard left or hard right are the only positions of zero phase , Center in stereo is kinda there but still subject to anomolies of stereo speakers room acoustics etc. That is why theatre sound developed the center speaker. to increase the clarity of dialog.

Um no. The center channel/speaker was incorporated to localize dialog at the screen in the front of room/theater, based on first arrival time heard by the audience.

So there are few real choices for pure punchy reproduction until you get to the bottom frequencies where the two speakers function together as a single array. (That will depend on the spacing of the speaker pair.) So to pan outside of the hard left and right creates "fuzz" in the mid and high frequencies, and the real question is how much fuzz do you want.

Stereophony is a bit more complex than that and mixing for film even more.

"Fuzz" is not an objective technical term.

JR

In a mix you can use that to contrast with the stuff that is in your face clear. Futher, adding of time delays choruses, Haas effect etc. can make mid and hi stuff clearer, but it is at expense of low and mid low frequencies which will add comb filtering that ruin any phase linearity. To add to the complication, people hear things differently. Some are able to isolate sounds, hearing around all the fuzz. To others it is just one big din. Others have one ear that hears things different from the other so they will favor one side over the other. So trying to come to a consensus is sometimes impossible. So unless there are individual speakers in all over the place - super Atmos surround? (Has anyone heard playback in George Massenbergs's Blackbird room) We make the best decision in the given situation with the given listeners.

 

[rh001]

Yucky is not an objective technical term.

JR

yucky is not meant to be objective. "Objective" only really matters if it's needed to make Money or as a defendable legal basis, for instance doctors or building engineers, if sued, will quite possibly be found not-culpable if they followed evidence procedures. But objectivity is not always required to make money as so many have proved like Sungha Jung and Yuzuru Hanyu, George Masenburg, Bill Putnum and the like who pursue a direction opposite of yucky (cuz they were probably born with good taste) and if not for these unobjective people life would be all yucky all the time.

 

[JohnRoberts]

yucky is not meant to be objective. "Objective" only really matters if it's needed to make Money or as a defendable legal basis, for instance doctors or building engineers, if sued, will quite possibly be found not-culpable if they followed evidence procedures. But objectivity is not always required to make money as so many have proved like Sungha Jung and Yuzuru Hanyu, George Masenburg, Bill Putnum and the like who pursue a direction opposite of yucky (cuz they were probably born with good taste) and if not for these unobjective people life would be all yucky all the time.

Objectivity matters in high performance circuit design. In marketing not so much.

JR

 

[Newmarket]

But this has been well established as usually being a question of pan law. Given that the centre:sides ratio is different by default in different consoles and DAWs, the same material may play back differently in them.
It is now less and less common, but 15 years back, it varied drastically, and was often not changeable in DAW options.

Beyond that, I also did think that before Pro Tools went 64 bit float, Logic sounded better to my ears.
My logic there is that there was more channel headroom (logic would read as if you had a peak, but not actually peak, given the essentially-infinite channel headroom, where PT would peak). This limited actual peaking to the output buss or plugins that didn’t have as much headroom.

Anyway, back to the actual question of preamp tone

None of that stuff should matter in a straight (ideally) bit for bit bounce.
fwiw I recall a conversation with Graham Boswell (Prism Audio) decades ago where he mentioned that you might be surprised at deviation introduced by some systems (best in mind this was mid 90s).

 

[Mr. Moscode]

Yucky is not an objective technical term.

JR

But still a term of art with varying degrees. Somewhere between "Oy Veh!", "Oh shit" and "WTF is wrong with that?", followed by a mad dash to the mute button.

 

[vintageampfixer]

On the topic of “punch and smoothness differences” in preamps, just adding to my earlier throw-in of headroom – because ability to handle the transient peaks when a saucepan lid’s dropped as well as having sufficient dynamic range to capture dialogue and the sound of a buzzing fly, etc., feels important to me – can I add phase distortion? To clarify, I mean the preamp’s phase response changing over the frequency range amplified. Any capacitor or inductor will by definition exhibit this characteristic and cause phase shifts which are typically greatest at the highest or lowest frequencies amplified.

 

[abbey road d enfer]

can I add phase distortion? To clarify, I mean the preamp’s phase response changing over the frequency range amplified. Any capacitor or inductor will by definition exhibit this characteristic and cause phase shifts which are typically greatest at the highest or lowest frequencies amplified.

So what? Phase-shift in minimum-phase systems (which most linear audio equipment are) is just a collateral of frequency response.
What do you think is the phase-shift of a circuit that passes 20Hz-20kHz within 0.5dB?

 

[vintageampfixer]

minimum-phase systems

That's my point: I'm not sure audio systems are "minimum-phase" if you compared the phase-shift at (say) 20Hz with that at 20kHz.

What do you think is the phase-shift of a circuit that passes 20Hz-20kHz

That's entirely dependent on the circuit design; a DC-coupled gain stage with no filtering shouldn't add any phase distortion but in most audio pre-amplifiers, stages are AC coupled and the coupling capacitors themselves introduce the effect with sequential AC coupled stages summing to the overall effect.

EQ or filtering, things will change again.

But it's not about absolute phase shift, per se. My last simply questions whether the variance in phase shift by frequency, which tends to be more notable at the extreme ends of the frequency response, could be related to “punch and smoothness differences” in preamps

 

[abbey road d enfer]

That's my point: I'm not sure audio systems are "minimum-phase" if you compared the phase-shift at (say) 20Hz with that at 20kHz.

Being minimum-phase has nothing to do with the amount of phase-shift at LF or HF.
Check it out on wiki.

That's entirely dependent on the circuit design; a DC-coupled gain stage with no filtering shouldn't add any phase distortion

No phase distortion at LF, because by consequence of being MP, there is no phase-shift at LF when the response extends to DC. In other words the LF cut-off frequency is zero Hz.

but in most audio pre-amplifiers, stages are AC coupled and the coupling capacitors themselves introduce the effect with sequential AC coupled stages summing to the overall effect.

That is correct.

But it's not about absolute phase shift, per se. My last simply questions whether the variance in phase shift by frequency, which tends to be more notable at the extreme ends of the frequency response, could be related to “punch and smoothness differences” in preamps

It would be if audition had the capability of evaluating phase.
Now I agree that a preamp with a frequency response of 20-20kHz has more phase-shift and probably less "punch" than one that goes 7Hz-60kHz. Why these values? Because it coincides with the 20hz-20kHz +0/-0.5dB I took as an example.

 

[vintageampfixer]

if audition had the capability of evaluating phase

Quite so for absolutes. But where there are phase shifts in some frequency ranges, I think they'd be audible.

a preamp with a frequency response of 20-20kHz has more phase-shift and probably less "punch" than one that goes 7Hz-60kHz

I don't follow ... a DC-coupled amplifier doesn't necessarily have a frequency response from 0Hz to whatever

 

[JohnRoberts]

On the topic of “punch and smoothness differences” in preamps, just adding to my earlier throw-in of headroom – because ability to handle the transient peaks when a saucepan lid’s dropped as well as having sufficient dynamic range to capture dialogue and the sound of a buzzing fly, etc., feels important to me –

Headroom or dynamic range is a well known objective metric.

can I add phase distortion?

Phase response or phase shift is predictable, phase distortion sounds like some non-ideal (?) behavior.

To clarify, I mean the preamp’s phase response changing over the frequency range amplified.

Are you talking about deviation from expected response, or normal phase shift wrt frequency.

Any capacitor or inductor will by definition exhibit this characteristic and cause phase shifts which are typically greatest at the highest or lowest frequencies amplified.

RLC phase shift is predicted by design equations.
====
[TMI- Ok I have shared this multiple times but not lately. Back in the 70s I sold a DJ mixer (kit) with two RIAA phono preamp inputs. I used a fairly conventional non-inverting topology pretty close*** to the National Semi app note. After the kits were out in the world I discovered some unexpected phase shift at 20 kHz. The phase shift (a couple tens of degrees at 20kHz) was not audible to me (with my sh__ for ears), but like I said it was not supposed to be there. Digging into a little deeper on my test bench I discovered that the phase shift was caused by an aluminum electrolytic capacitors ESL (equivalent series inductance). This is also a well known phenomenon. In my circuit the 22uF electrolytic capacitor was in series with a 360 ohm gain resistor for the NI RIAA preamp. This delivered a nominal -3dB @ 20Hz LF pole. What was not expected was that the amplitude response at 20kHz correctly followed the RIAA 75uSec roll off pole, but with an unexpected 20-30' phase shift from the aluminum electrolytic's ESL. Replacing the 22 uF aluminum electrolytic caps with 22 uF tantalum, tightened up the 20kHz phase response.

I repeat I could not hear a difference, but decided to experiment with a bunch of my kit customers. I mailed about ten of them pairs of 22uF tantalum caps and asked they to replace the caps first in one preamp and then do some simple listening tests comparing them. I only heard back from a few customers. Most claimed that they heard an improvement, but there was likely some expectation bias in those uncontrolled listening tests. One customer told me that he didn't try the tantalum caps because "everyone knows" that tantalum caps have lousy dielectric absorption. (he shall remain nameless but was a known guy back then and is now RIP).

FWIW these aluminum caps from back in the 70s were nothing like today's much improved electrolytic caps.

====
The lesson I took from that bench work was not to use aluminum caps in series with relatively low impedance gain setting legs. Even though I couldn't hear it I could measure it, and didn't like it.

Do not interpret this as some kind of support for your audibility of normal phase shift.

JR

*** for TMI on top of my TMI veer... the improvement I made to the nat semi app note was regarding stabilizing the op amp. For improved performance they used a decompensated op amp that was only stable for closed loop gains of several X. For accuracy the RIAA EQ curve involves a real pole at 75uSec that in theory keeps rolling off forever. The non-inverting topology starts out with one strike against it because the NI topology can not fall below unity gain, but a second strike comes from the Nat Semi Ap note that mandated a several X closed loop gain on top of that for stability. My improvement was to add the extra gain resistance for stability, but to take my audio feed from the top of the caps in feedback network. This gave me a more accurate HF RIAA response (except for the NI unity gain zero). Since I was feeding relatively high impedance nodes inside my DJ mixer there was no problem from tapping the signal where I did. This DJ mixer was published back in 1978. I designed several RIAA preamps after that and never used the NI configuration again. All my later RIAA preamps continued the 75uSec roll off up to light.

 

[kags]

I normally work in Pro Tools HDX, but I receive a fair amount of files in Logic Pro (v 10.7.x), so I tend to listen a lot to Logic sessions to decide whether I should continue working on a client's project in Logic or if I should export the audio over to Pro Tools. I tend to only remain in Logic if the project is extremely close to sounding finished. In any case, I am 100% sure that when I bounce a mix out of Logic a few times the mixes sound different from each other. They do not null 100% when imported to another DAW. I have a sense that it's mostly time-based effects that play differently with each pass, but I can't be convinced that all the automation plays correctly each time I play the session back. Usually the differences that I hear between mix passes are not different enough that I prefer one vs. another, but it really bothers me that this even happens. In my experience, it does not happen in Pro Tools, Reaper, Studio One, or Cubase.

Years ago I spoke with Dave Hill (CraneSong) about his plugins being available only for Pro Tools (TDM then AAX), and he said he couldn't trust that VST or AU plugins performed consistently. He trusted Pro Tools DSP processing to be "bit-perfect" and reliable and repeatable. I imagine other DAWs have similar integrity in 2024, but who knows? Whenever I have colleagues who mention that different routings in Pro Tools sound different to them, we can always find some difference in level or processing that causes the difference, even if the difference is < 1 dB.

When I'm programming a song, I'm not too concerned about the minute randomness of a DAW's behavior, but in mixing and mastering I don't think it's acceptable.

 

[ccaudle]

I'm not sure audio systems are "minimum-phase"

Multi-driver speakers are not. Any reasonable preamp design is minimum phase. A circuit with all-pass filters would not be (e.g. some types of phasor or chorus circuits).

where there are phase shifts in some frequency ranges, I think they'd be audible.

Does that include taking a step forward or backwards?

 

[ruffrecords]

That's my point: I'm not sure audio systems are "minimum-phase" if you compared the phase-shift at (say) 20Hz with that at 20kHz.

Unfortunately that is a pointless exercise and shows just how little most people understand about the relationship between frequency, phase and delay.

Simply put. all systems will have a delay. A delay will cause a phase shift which increases proportionally with frequency and the output will be the same as the input. In other words, you should not be aiming for minimum phase but linear phase. If you cannot measure a phase difference between 20Hz and 20Kz then something is wrong.

Cheers

Ian

 

[abbey road d enfer]

you should not be aiming for minimum phase but linear phase.

It's not a matter of aiming.
Minimum phase is just a consequence of systems that don't have multiple paths with different delays, which is the case of many building blocks, like preamps and amps.
Linear phase implies a pure delay, which is most of the times not the case. Any building block has at least one HF pole and often a LF one, which results in phase-shift at HF and LF, accompanied with associated roll-off.
In practice linear phase is always an approximation, with a specific domain of validity. Phase linearity is spec'd as +/- x degrees in a frequency range.

Simply put. all systems will have a delay.

Yes, but in practice this delay is far from being constant.

 

[abbey road d enfer]

In any case, I am 100% sure that when I bounce a mix out of Logic a few times the mixes sound different from each other.

What do you mean "bounce a few times"?

They do not null 100% when imported to another DAW. I have a sense that it's mostly time-based effects that play differently with each pass, but I can't be convinced that all the automation plays correctly each time I play the session back.

Again, are you sure that the pan laws are exactly the same in PT and Logic?
Time based FX do not sound the same because there is a degree of randomness in the algorithms, particularly reverbs, where the tails are randomly modulated.